Specificity in scikit learn

Question

I need specificity for my classification which is defined as : TN/(TN+FP)

I am writing a custom scorer function :

from sklearn.         


        

Answer 1

Remembering that in binary classification, recall of the positive class is also known as “sensitivity”; recall of the negative class is “specificity”, I use this:

unique, counts = np.unique(y_test, return_counts=True)

for i in unique:
    score = precision_score(y_true, y_pred, labels=unique, pos_label=i)
    print('score ' + str(i) + '  ' + str(score))